A contrastive analysis of the present progressive in French and English 1 Astrid De Wit, Adeline Patard, Frank Brisard

Transcription

1 A contrastive analysis of the present progressive in French and English 1 Astrid De Wit, Adeline Patard, Frank Brisard [Note that this version is a pre-final draft; please consult the following printed version for referencing: De Wit, Astrid, Adeline Patard & Frank Brisard (2013), A contrastive analysis of the present progressive in French and English. Studies in Language 37(4): ] In this study, we address the semantics of the present progressive constructions in French and English by looking into their present-day uses and their diachronic evolution. Corpus data show that both constructions are frequently used in contemporary English and French to stress the atypical nature of situations. This suggests that these constructions share an epistemic core meaning, which we define as contingency in immediate reality. However, in terms of concrete usage types which elaborate this meaning in context, the two progressive constructions differ significantly: the French progressive occurs in fewer types of context than its English counterpart and it is, overall, less frequently used and not obligatory (e.g., for referring to present-time events, as in English). We argue that these differences can be systematically related to the different diachronic evolutions that have shaped the present-tense paradigms more specifically, the respective aspectual values of the simple present-tense markers in both languages. Keywords: present progressive, simple present, English, French, aspect 1. Introduction This study addresses the semantics of the French present progressive in comparison with its English counterpart. It is well-known that the French être en train de + V-inf construction, like many other European progressives, is less grammaticalized than the frequently used English be + V-ing form, but the descriptive details of and the reasons for these differences are hardly ever dealt with (for notable exceptions, cf. Bertinetto 2000, Lachaux 2005 and Author 1 & Author ). In order to look into these issues more thoroughly, we have conducted a study of the 1 The research carried out for this paper has been financed by the Research Foundation Flanders (FWO) (Projects 4740 and 5295). The authors wish to thank Peter Petré, Naoaki Wada, two anonymous reviewers and the audiences of Chronos 10 (Birmingham, April 2011) and of the 45th Annual meeting of the Societas Linguisticae Europeae (Stockholm, 28 August 1 September 2012) for their useful comments on earlier presentations of the material in this paper. 1

2 present-tense uses of the progressive in various corpora of spoken French (Elicop 2, CFPP and CLAPI 4 ). The different uses of the French progressive form have systematically been compared to those of the English present progressive, as attested by Author 1 & Author 3 (forthcoming) in the Santa Barbara Corpus of Spoken American English, Part 1 (SBC; Du Bois et al. 2000). This comparison shows, among other things, that the English present progressive is more frequently (and, in some contexts, obligatorily) used and occurs in a broader range of contexts than its French counterpart (cf. Author 1 & Author ). At the same time, the two constructions turn out to share quite a few usage types as well. In fact, the uses of the French present progressive appear to constitute a proper subpart of those of the English present progressive. The main goal of this paper is to account for these hitherto unexplained observations. We will argue, more specifically, that the French and the English present progressive share the same core meaning in that they both indicate epistemic contingency in current reality (Author 1 & Author ) and that the present-day differences in use between the two constructions are due to the different diachronic evolutions that have shaped the presenttense paradigms in the two languages under consideration. In Section 2, we will summarize the most important results from the corpus work presented in Author 1 & Author 3 (forthcoming) and Author 1 & Author 2 (2013). This includes an overview of the various uses, temporal and modal, of the English and the French present progressive. A comparative analysis of these corpus data will be presented in Section 3. Section 3.1 is devoted to the modal uses and the modal core meaning of the present progressive in both French and English. In Section 3.2, then, we tackle the differences in use between the two constructions. In Section 3.2.1, we outline the different diachronic pathways of the English and French progressives and we address the relevance of these different histories for the distribution and use of the simple and the progressive constructions in the present-tense paradigms of the two languages, thereby also introducing evidence from diachronic evolutions in the tense and aspect systems of other Germanic and Romance languages (especially, but not solely, Dutch and Spanish). It will be argued, more specifically, that, under influence of the heavily grammaticalized progressive, the simple present in English has become a perfective present tense, with a fairly restricted set of usage types, while the French simple present is aspectually ambiguous. Next, in Section 3.2.2, we explain why the French present progressive hardly ever 2 Cf. [ ] 3 Cf. [ ] 4 Cf. [ ] 2

3 expresses duration or iteration and resists a futurate interpretation. In Section 4, finally, we submit our conclusions. Before we embark on our study, however, we need to explain why we are only focusing on the present progressive, thus excluding past-tense progressive uses from our analysis. We have, in fact, various reasons for doing so. First, as we will show in Section 3, there is a strong link between the use of the present progressive and the properties of the speech event, especially in English. That is, we will argue that the progressive is indispensable in English to align dynamic situations (i.e., events) with the time of speaking, while there is no such strong requirement in the past. For French, an indication of the differentiation/specialization between the present and the past progressive comes from Lebas-Fraczak (2010), who shows that some previous accounts of the semantics of être en train de + V-inf hold for the present but not for the past, as these two domains involve different modal interpretations. Data from other languages (cf., e.g., Güldemann 2003 on (mostly) Bantu languages) equally show that the notable association of the progressive with modal connotations appears to be particularly relevant for the present domain, rather than for the past. These observations indicate, in our view, that the semantics of the present and past progressive constructions need to be studied separately. This does not entail, however, that evolutions in the past-tense paradigm have never been of any influence for the present-tense constructions under consideration. For instance, we will argue in Section that the aspectual ambiguity of the simple past, which is characteristic of English (in contrast with Dutch and German) may have enhanced the grammaticalization of the progressive, first in the past and then extending to the present. The French progressive, on the other hand, enters in competition with another, much more highly grammaticalized aspectual construction in the past paradigm, i.e., the imparfait a competition that has undoubtedly influenced the evolution of the present progressive as well (cf. Section ). 2. Usage types of the English and the French present progressive In this section, we summarize the results of the corpus studies on the use of the English and the French present progressive, presented in, respectively, Author 1 & Author 3 (forthcoming) and Author 1 & Author 2 (2013). In the SBC (part 1), consisting of 50,000 words, a total of 339 examples of be + V-ing have been collected, while in the French corpora, together containing approximately 2 million words, only 191 examples of être en train de + V-inf could be counted, which boils down to merely 4.78 occurrences per 50,000 words. This clearly indicates that speakers of English much more frequently use the present progressive than speakers of French. 3

4 Table 1 lists the various usage types arrived at and their frequencies in both languages. Each of the attested examples has been categorized as instantiating a particular usage type on the basis of (i) contextual cues, such as adverbs, which elaborate meaning elements present in the configuration of the progressive, and (ii) the difference in meaning, according to native speakers, generated by replacing the progressive with a simple present, as we will illustrate in the relevant contexts below. Table 1: Usage types and frequencies of the English and the French present progressive 5 English French Absolute Percentage Absolute Percentage number numbers Current 93 27, ,16 ongoingness Historical 71 20, ,28 present Temporary 23 6, ,24 Validity Duration 16 4,72 1 0,52 Incompletion 2 0, ,90 Futurate 46 13, Iteration 17 5,01 2 1,05 Habitual 43 12, ,42 Interpretative 28 8, ,42 Total In the remainder of this section, we will first discuss the aspecto-temporal uses of the present progressive in French and English, i.e., those uses in which the construction is used to refer to events that are going on at the time of speaking. Then we turn to the modal usage types and modal connotations that often accompany the use of the present progressive in both French and English. With these uses, aspecto-temporal meanings of current ongoingness are highly downplayed, if present at all. As we will show in Section 3.1, such modal meanings reflect the present progressive s core semantics (instantiated in any of its uses in the two languages), which we will define as epistemic contingency in immediate reality (in line with Author 1 & Author and Author 1 & Author 3 forthcoming). Table 1 shows that the English present progressive examples are more evenly spread across the range of possible usage types than instances of its French counterpart, which is most often used to express Current Ongoingness. Yet in English, too, Current Ongoingness constitutes, on 5 Tables 1 and 2 have been adopted (with some minor changes) from Author 1 & Author 2 (forthcoming). Names of usage types have been capitalized. 4

5 the whole, the most prototypical (i.e., most frequently occurring) category of use of the present progressive. In the examples belonging to this category, the progressive is used to zoom in on a dynamic situation that is going on at the time of speaking, without any further specifications, like, for instance, suggesting (unusual) duration (cf. also Bertinetto 2000: 565) see examples (1) and (2) for illustrations: (1) So I stopped the car and they said: What are you doing? I said: Oh, I gotta tighten this wire here. (SBC007) (2) Je suis en train de chercher le nom mais je l retrouve plus. (cfpp2000) I m looking for the name but I can t find it anymore. In English, the use of the progressive is obligatory in such contexts, while in French it is optional (and sometimes even unnatural or impossible). 6 For instance, replacing the progressive by a simple form would be downright ungrammatical in (1), but perfectly possible in (2), in which case the only difference in meaning would be that a simple-present tense construal of the same event (je cherche le nom) does not necessarily imply that the denoted mental activity is actually taking place at the time of speaking (it could, for instance, also be a habit on the part of the subject). The other aspecto-temporal usage types of the progressives in the two languages may be regarded as more specific instantiations of the category Current Ongoingness: they involve events that are ongoing at the time of speaking, but extend or elaborate this meaning by slightly changing it or adding something to it. For instance, in English in particular, the present progressive may also be used to refer to (actually) past events that are construed as occurring in the present cf. (3) for an illustration of such a Historical Present use: (3) Two weeks ago I m watching TV, and David Horowitz is going to have this former car radio thief on? (SBC006) In French, too, examples of Historical Present uses of the present progressive can be attested, although they seem to occur more rarely than their English counterparts: 6 Do-Hurinville (2007: 8) points out, for instance, that être en train de + V-inf does not normally combine with en ce moment ( at this moment ). 5

6 (4) Alors voilà, c'est l' histoire d' un homme euh qui a tué une poule d'eau et qui est en train de la déplumer et alors il avait pas le droit et malheureusement pour lui il y a le garde champêtre qui arrive (Elicop) Well, then, this is the story of a man who has killed a hen and who is plucking it and so he didn t have the right and unfortunately for him the village policeman arrives Both the English and French present progressive may be used as well to refer to temporarily valid events, as in (5) and (6): (5) So they re kind of suffering that from that this year. Not having that on there. (SBC006) (6) On a travaillé pour le le temple protestant de Monaco. On on est en train de faire des vitraux pour euh Caen [ ] je vais faire des vitraux prochainement à la Madeleine à à Saint-Maxime à la Sainte Beaume dans le Var. (Elicop) We ve been working on the the protestant temple of Monaco. [At present] we we re making stained glass for Caen [ ] soon I m going to make stained glass for the church of the Madeleine in Saint-Maxime in the Sainte Baume mountains in the Var. Given the explicit indications of temporariness (this year in (5) and the reference to preceding and following events in (6)), using the simple present in these example would sound less natural. Note, further, that (5), like quite a few other examples featuring a progressive (cf. Table 2 below), involves a sense of tentativeness, as indicated by the presence of kind of. Next, in French as well as in English, events may be explicitly presented as incomplete, as in (7) and (8): (7) [ ] Non, pas encore. Ah oui, euh, quand ils vont avoir fini la couverture là, euh Ils sont en train de me finir l'autre chantier, alors les plombiers vont s'en aller de là-bas. (Elicop) [ ] No, not yet. Oh, yes, when they re going to have the covering finished there, uhm They are clearing the other construction site, so the plumbers will go from there. (8) So now Fletcher s starting to realize what s going on. (SBC006) It is easier to identify instances of the category of Incompletion in French than in English. This is due to the fact that, in English, every telic verb is as a rule given a progressive construal and 6

7 thus by definition presented as incomplete. On the other hand, être en train de + V-inf is, in some examples (such as (7)), more clearly employed only for the expression of incompletion, since its use is not grammatically obligatory (cf. Author 1 & Author for a more elaborate discussion). 7 Events in the progressive may also be interpreted as having (prolonged) duration. According to Bertinetto (2000), the French progressive can never express duration, and it has never done so either, in contrast with other European progressives which have all gone through a durative stage. Our corpus data show that such durative uses of être en train de + V-inf are indeed very restricted but not completely excluded, as we have been able to find one example: (9) Ça fait un an que je suis en train de faire un truc qui est INCROYABLE. Je sais pas si tu te rends compte. (Clapi) For a year I ve been doing this INCREDIBLE thing. I don t know if you realize. This and other examples of the type ça/cela fait [noun phrase indicating a particular duration] que found on the internet, can take être en train de + V-inf and thus undermine Bertinetto s (2000) claim. Yet it should be noted that these durative uses always involve prolonged situations that are often in a way surprising or sometimes even irritating to the speaker. If (9) featured a simple present rather than a progressive, such a subjective connotation would be less conspicuously present, if at all. In English, too, durative uses, which are more commonly attested, often feature a sense of surprise or irritation, as in (10): (10) RICKIE: And then the whole time under here, he d just look. I mean, he looked so hard that it was, like, burning. [ ] REBECCA: So then, and then, he sort of pulled the paper aside, and he s still staring at you? (SBC008) A striking difference between the French and the English progressive is the fact that the former does not sanction a futurate interpretation, while this is quite natural in English, as in (11): 7 The category Incompletion is the only category whose meaning is rather intertwined with the actional semantics of the verbs involved. All the other categories (both in French and in English) do not exhibit any preference for a specific type of lexical aspect, according to our corpus analyses. There are, for instance, no indications that, say, Current Ongoingness is more often attested with activity predicates than with accomplishments in comparison with other categories. 7

8 (11) A: Rosenblum is coming back though, isn t he? B: I don t know what he s gonna do. (SBC006) These futurate uses of be + V-ing are especially attested in contexts in which the speaker is quite confident, yet not fully certain, about the future occurrence of the denoted event (because, for instance, it has been arranged in the present). Iterative and habitual interpretations involve the repetition of an event either rapidly and within a single, short time span (Iteration, cf. (12) and (13)) or within a larger time span (Habitual, cf. (14) and (15)). These uses also turn out to be much more common in English than in French iterative uses, in particular, are very rare in French. Note that habitual uses of the present progressive often feature a sense of atypicality, surprise or irritation, in English as well as in French: (12) [ ] Cause their hips are beating up against you, you know. Like that [slapping], fast. (SBC002) (13) Le petit garçon euh saute pour le [son chien] pour le chercher toujours euh il est toujours en train de l'appeler et euh i va enfin il va à la forêt. (Clapi) The little boy is jumping, looking for him [his dog] still uhm he s still calling him and finally he goes to the forest. (14) Everywhere we ve been, in the past several years, everybody s talking about how, the weather just isn t normal. (SBC003) (15) Dès qu'ils ont une place apparemment ils sont en train d'mettre des des immeubles de bureaux. J'sais pas si vous avez remarqué. (CFPP 2000) As soon as they have space apparently they re putting in office buildings. I don t know if you ve noticed. In both French and English, the progressive may in addition be used to re-interpret a situation (cf. Ljung 1980 for an in-depth discussion of these so-called interpretative uses). Such uses typically involve contexts in which the speaker reformulates a previously mentioned reference to an event or refers to an event the (underlying) meaning of which she does not consider straightforward to interpret, as in (16) and (17): 8

9 (16) [In a discussion between a professor and his students about the discourse of civil rights activist Jesse Jackson:] He s smart, he talks about minorities. But he s really talking about African Americans. (SBC012) (17) A: Le quartier a des tas de liens, ne serait-c' que par l'école et les paroisses, hein B: Absolument, c'est vrai A: Ce que vous êtes en train de dire c'est qu'il y a une présence catholique forte encore. (CFPP2000) A: In the neighborhood there are a lot of bonds, if only through school and the parishes, aren t there. B: Absolutely, that s true A: What you re saying is that there s still a strong catholic presence. In example (16), the speaker is reporting on a habitual action of the subject (Jesse Jackson) and there are no aspecto-temporal differences between the first reference (rendered in the simple present) and the second one (given a present-progressive construal). In the French example in (17), too, the present progressive is used to reformulate an observation, rather than to indicate ongoingness in the present. In cases such as these, the construction is purely called upon to make the event stand out and to emphasize its actual nature, i.e., it is used for modal (epistemic) rather than for aspecto-temporal reasons (cf. the presence of the adverb really in (16)). This brings us to the following central observation: the French and English examples often carry subjective notions of surprise and irritation (cf., e.g., examples (9), (10), (14) and (15)), tentativeness (cf. (5)) and intensification (cf. (12)) (cf. the frequencies in Table 2 below). In some cases, conveying such a subjective meaning appears to be the main, if not only, reason for using the progressive instead of the simple form (i.e., there is no aspecto-temporal reason for doing so). Examples (18) and (19), for instance, involve habitual situations and it would, therefore, be grammatically correct to use a simple present in both the English and the French sentence, yet this would sound pragmatically anomalous. In (18), the speaker presents the situation in a more intensified manner (cf. the presence of really really), while (19) features a sense of irritation (as is also reflected in the use of ce it rather than ils they in subject position). Apparently, subjective expressions such as these typically elicit the use of the present progressive rather than the simple present, which would present the denoted situations in a more neutral fashion. 9

10 (18) I always have somebody that really knows what they re doing for the horses that I m really really using. (SBC001) (19) En ce moment, hein, à notre époque beaucoup les cadres les fils de bourgeois qui font les [ ] hein c'est tout le temps en train de voyager ces gens-là hein. (Clapi) At the moment, huh, these days a lot the executives the sons of bourgeois that do the [ ] huh they re [lit. it s] travelling all the time those people huh. In many other uses, aspecto-temporal and modal motivations for using the progressive seem to go hand in hand. In (20), for instance, the use of the present progressive is grammatically obligatory as the denoted situation is ongoing in the present, but it also naturally ties in with the atypical nature of the situation: (20) And it s a killer on your back cause you re standing like this. (SBC002) Since aspecto-temporal motivations for using the progressive are almost always present in the background (especially in English, in which the use of be + V-ing is obligatory with dynamic situations in the present), we do not regard surprise, tentativeness, irritation and intensification as purely modal (in the sense of non-aspecto-temporal ) usage types (such as the Interpretative use), but rather as subjective expressions that typically collocate with aspectotemporal uses of the present progressive (since they, as we will demonstrate in Section 3.1, reflect the construction s core meaning of contingency in immediate reality ). Table 2 lists the number of examples that feature a sense of surprise, tentativeness, irritation or intensification in both English and French (note that these examples have already been classified as instantiating a particular usage type in Table 1). Table 2: Modal expressions in English and French English French Absolute number Percentage (total of 339) Absolute number Percentage (total of 191) Surprise 62 18, ,33 Tentativeness 11 3, ,90 Irritation 36 10, ,33 Intensification 50 14, ,37 The results of the corpus studies summarized here yield a number of interesting results, which in turn give rise to the research questions that will be dealt with in Section 3: 10

11 (i) (ii) The French as well as the English present progressive relatively frequently involve modal connotations, which would not be present if the simple present were used instead. 8 How do these tie in with the aspecto-temporal uses of the progressive constructions and with their core semantics, more generally? These questions will be discussed in Section 3.1, in which we argue that the present progressive in both French and English expresses epistemic contingency in immediate reality at the most basic level of analysis (Author 1 & Author ). The English present progressive can express a broader variety of aspecto-temporal meanings than its French counterpart, which is more or less limited to expressing pure Current Ongoingness (although even in that context, its use is not obligatory). For most other categories 9, the use of être en train de + V-inf is quite and sometimes even heavily (cf. Duration and Iteration) restricted, and future-time reference is excluded with this construction. In other words, the usage types of the French present progressive constitute a subpart of those of the English present progressive. The use of the latter construction is, moreover, much more entrenched: not only because it is grammatically obligatory with dynamic verbs, but also in view of its frequency of occurrence (as we have noted at the beginning of this section). Our findings thus confirm those of Lachaux (2005), whose study of English texts and their French translations shows that, in a great many cases, the English progressive cannot be translated by means of its French equivalent (i.e., être en train de + V- inf). These weak functional limitations, the high frequency, and obligatorification of be + V-ing can be regarded as indicative of its high degree of grammaticalization. While these observations are not surprising as such (since the English progressive is known for its notable degree of grammaticalization), they do bring up a couple of questions that have thus far not been tackled in any comprehensive analysis: a. Why is the English present progressive more grammaticalized than the French one? As we will see in Sections 3.2.1, there are various historical explanations for this high degree of grammaticalization of be + V-ing (Section ) and for the comparatively low degree of grammaticalization of its French 8 For English, this opposition between the simple present and the present progressive can of course only be made apparent in those contexts in which the use of the simple present would be grammatical in the first place, such as (14), (16) and (18). 9 The category of Incompletion is not typical in this respect, in that, as we explained in Section 2, it is easier to identify instances of it in French than in English. 11

12 counterpart (Section ). In this section on diachrony, we will also integrate a discussion on the (evolution of the) aspectual semantics of the English and French simple present constructions. Assuming that the communicative need for expressing progressive aspect is equal in the two languages (Mortier 2008: 5-6), we may predict that this function is partially fulfilled by another construction in French. In Section , we will show that this other construction is the French simple present tense, which can express imperfective (as well as perfective) aspect. It is thus distinct from the English simple present, which has, under influence of the independently developing present progressive, specialized into solely marking perfective aspect (Section ). b. Why do some meaning categories in particular resist expression by means of the progressive in French, while they quite naturally occur with the progressive in English? To account for the restrictions on durative and iterative uses in French, we will again refer to diachronic differences between the English and the French progressive periphrases. That is, as will be demonstrated in Section 3.2.2, the more dynamic origin of être en train de + V-inf lends itself less naturally to relatively static interpretations of prolonged duration and iteration than the locative origin of be + V-ing. The absence of futurate readings for the present progressive in French, on the other hand, will be related to the comparatively wide array of possible uses of the aspectually ambiguous simple present, which consequently constitutes the default and only dedicated present-tense construction to be used in future-time contexts, as opposed to its more restricted (i.e., solely perfective) equivalent in English. As we will indicate, a futurate use of the present progressive, commonly attested as it may be in English, actually turns out to be fairly exceptional from a cross-linguistic perspective. 3. Semantic analysis of être en train de + V-inf, compared to be + V-ing: Synchrony and diachrony 3.1. Modal core meaning of the two present-progressive constructions Many analyses of the semantics of the English progressive focus on purely aspecto-temporal notions: it has been argued to express temporariness (Mindt 2000), duration (Palmer 1989), limited duration (Quirk et al. 1985; Leech 2004), imperfectivity (Declerck, Reed & Capelle 12

13 2006), progressiveness (i.e., ongoingness at a particular reference point) (Huddleston & Pullum 2002: 162; Tharaud 2008), temporal framing (Jespersen 1931) and incompletion (Palmer 1989; Leech 2004). Yet we have seen that the progressive appears in a number of idiosyncratic contexts too (Comrie 1976: 38), in which there does not seem to be any aspecto-temporal motivation for introducing the construction: it can be used to refer to future situations (often arranged in the present), to re-interpret a situation and to express some emotional evaluation (irritation, tentativeness ). Especially the latter two functions are often neglected in standard analyses of the core meaning of the English progressive notable exceptions are Adamczewski (1978), Ljung (1980), Wright (1994), Rydén (1997), Williams (2002) and Tharaud (2008). Even though some of these authors try to provide unified accounts of the various uses of the English progressive, these are not entirely satisfactory since certain usage types are either neglected (Rydén 1997, for instance, explicitly disregards the futurate use of the progressive in his analysis) or excluded from the semantics of the construction proper (because, for instance, they are argued to be derived from the context rather than from the meaning of the progressive, cf. e.g. Tharaud 2008). In other cases (notably Adamczewski 1978 and Williams 2002), the suggested accounts (comprehensive though they may be) do not succeed in proposing one basic semantic schema that systematically covers and links all of the attested usage types (see Author 1 & Author 3 forthcoming for further discussion). The French progressive has understandably attracted less attention than its more frequently used English counterpart, yet there are some interesting analyses of its core semantics. Do- Hurinville (2007), for instance, suggests dynamism as the diachronic core meaning of the French progressive, i.e., the meaning that has always been part of the semantics of the progressive as well as of its component parts. As we will argue in Section 3.2.2, this dynamic component in the progressive s semantics has substantially influenced its evolution and presentday semantics. Some other analyses of the core value of the French progressive, viz. those of Franckel (1989) and Lachaux (2005), explicitly include the modal expressions associated with the construction s use. The main point of these proposals which are to be situated within the enunciativist tradition, which focuses on the pragmatic, inter-speaker functions of grammatical devices is that the progressive marks a situation as standing out, i.e., as divergent with respect to what is expected, in a very broad sense (cf. mise en relief, Lachaux 2005). Franckel (1989: 78) offers the following example, which nicely illustrates that the present progressive more naturally refers to atypical events (21a) than to situations that are expected to occur (21b): 13

14 (21) a. Il est en train de lire de travers. He is reading the wrong way. b.? Il est en train de bien lire. He is reading well. Leeman (2012) further elaborates on Franckel s (1989) analysis, suggesting that the meaning of discrepancy (typically involving a negative evaluation) is compositionally derived from the meaning of the preposition en and the verb traîner ( to pull, drag ). These accounts of the semantics French progressive tie in closely with the semantic analysis of the French and English present progressive put forward in Author 1 & Author 2 (2013), who claim that, in both languages, the present progressive has a modal (epistemic) meaning at the core of its semantics. It is argued to indicate, more specifically, that a situation has a contingent status within the speaker s conception of present (immediate) reality. This means that, even if an event is actually ( really ) happening at the time of speaking, the fact that it is happening now and how it manifests itself could not have been predicted or expected (cf. also Author 1 & Author 3 forthcoming). In other words, the event is not construed as part of the structural world knowledge of the speaker (Goldsmith & Woisetschlaeger 1982; Brisard 2002), i.e., it is phenomenal (observable as (if) developing before the conceptualizer s eyes). Example (21) constitutes a good illustration of this opposition between structural and phenomenal situations in French: (21a) represents a phenomenal, contingent situation, whereas (21b) involves a situation that constitutes a more structural part of the speaker s reality (since the subject is presumably expected to know how to read). An English example of the structural/phenomenal opposition is suggested in Goldsmith & Woisetschlaeger (1982: 81): (22) a. The engine isn t smoking anymore. b. The engine doesn t smoke anymore. As Goldsmith & Woisetschlaeger (1982: 81) point out, it would be distinctly odd to utter sentence (22b) in a context in which the speaker was not expecting the engine to be repaired. In such more phenomenal circumstances, the use of (22a), which merely describes what is happening, is much more appropriate. Conversely, if the speaker does expect the engine to have stopped smoking (because, for instance, she is confident that she has repaired it), (22b) does appear appropriate. The use of the simple present then reflects that, in the normal course of events, i.e., given a certain knowledge of the structure of the world, the engine indeed does not 14

15 smoke anymore after it has been repaired. The meaning of contingency characteristic of the present progressive naturally gives rise to connotations of surprise, intensification, tentativeness and irritation, which, therefore, frequently accompany the use of the present progressive, and to the interpretative meanings, as these refer to situations that are regarded as exceptional (and thus not straightforward to interpret, offending etc.). If this modal analysis in terms of contingency in immediate reality is valid for all the uses of the progressive (in French and in English), then one might wonder how it applies to the apparently purely aspecto-temporal uses with no such additional modal connotations. The answer, Author 1 & Author 3 (forthcoming) argue, resides in the backgrounded nature of the boundaries of an event that is imperfectivized by means of the progressive. The following figures, adopted from Langacker (2001: ), illustrate how this works. First, the progressive, as an imperfective marker, creates an internal perspective (an aspectual immediate scope, ISA; morphologically marked by ing in English and by en train de in French) on a dynamic situation, which is by definition bounded in time (events always have some limit (Langacker 1987: )) and, often, heterogeneous. This progressive configuration is illustrated in Figure 1. IS A MS t Figure 1: Internal perspective imposed by the progressive 10 As a result of this internal viewpoint, the boundaries of the event are backgrounded, i.e., they are put in the configuration s maximal scope (MS). 11 At the same time, that part of the situation that is within the progressive s scope (ISA) has been homogenized, moving the focus away from its evolution towards a final boundary and concentrating on its mere development or passing in time. Such imperfectivized dynamic situations, as well as states, are contractible (Langacker 2001) or in interval-semantic terms they have the subinterval property (Bennett and Partee 1978). This means that any random segment of the situation is representative of the situation in its entirety a property that is not shared by perfectively viewed dynamic situations. By virtue 10 For ease of representation, we will refrain from explicitly indicating the prototypical heterogeneity of events in the subsequent figures. 11 When stative verbs are given a progressive construal, these boundaries are created i.e., the denoted states are coerced into something more dynamic (and temporary). 15

16 of this characteristic, states and imperfectivized events can be fully verified at the time of speaking, and they are therefore, unlike perfectively viewed events, compatible with the present tense, which implies full and exact coincidence between the event and the time of speaking (Langacker 1991) (in Section 3.2.1, we will propose a more elaborate analysis of the interaction between the present tense and types of lexical and grammatical aspect). This alignment of an event with the time of speaking is depicted in Figure 2, in which a second scope, that of the present tense (the speech event is indicated by the squiggly line and the temporal scope is indicated by IST), is superimposed over the first immediate scope. This represents the fact that the speaker denotes, by using a present tense, that part of the situation that overlaps with her description of it. IST IS A MS t Figure 2: Configuration of the present progressive (in French and English) Given this definition of the present progressive, we can assume that, in each of its uses, backgrounded boundaries are by necessity implied (even with states; cf. Footnote 11) (cf. also Leeman (2012) on the boundaries imposed by être en train de + V-inf). The presence of these boundaries, together with the progressive s internal (i.e., imperfective) perspective, directly reflects the meaning of epistemic contingency that is, the temporal configuration depicted in Figure 2 results in a less than complete view offered on the situation and thus in less than certain knowledge on its further development and possible culmination Grammaticalization of être en train de +V-inf, compared to be + V-ing Different histories, different aspectual values In the previous section, we have postulated an important semantic parallel between the French and the English present progressive, i.e., they share the same basic, crucially modal meaning 12 Cf. Dowty s (1979) analysis of the interaction between telic events and the progressive (known as the imperfective paradox ). 16

17 (as is reflected in their frequent association with modal connotations related to the meaning of contingency). Despite this parallelism at a schematic, more abstract level, the constructions differ at a more specific level, i.e., in actual usage: even though they both prototypically refer to currently ongoing events, the French present progressive is used less frequently, it is not grammatically obligatory with dynamic verbs and semantically less differentiated. The explanation for these differences resides, in our view, in the different diachronic evolutions that have shaped the present-tense paradigms in the two languages at hand History of the present progressive and the simple present in English Let us start with the diachrony of be + V-ing and its ancestor, beon/wesan + V-ende. 13 Given the remarkable degree of grammaticalization of the English progressive from a cross-linguistic perspective (cf., e.g., Comrie 1976: 32), quite a few studies have been devoted to its diachronic development in order to shed more light on this aspect. The most relevant comprehensive descriptions, on which we are basing our diachronic analysis, are found in Scheffer (1975), Wright (1994), Ziegeler (1999), Núñez-Pertejo (2004) and Kranich (2010). Old English inherited from Proto-Germanic an aspectual system that is reminiscent of a type of aspect-marking which is also attested in present-day Slavic languages: often, imperfective verbs (i.e., verbs with an atelic, durative or stative, meaning) were turned into perfective (i.e., telic) verbs by means of prefixing, notably by ge-, but also by, for instance, a-, be-, on- and for- (Núñez-Pertejo 2004: 66-67; Scheffer 1975: ). 14 Consequently, the verbs that remained unmarked were commonly interpreted as being imperfective (typically atelic and durative). As explained in detail in Brinton (1988), this association between perfectivity and prefixed verbs and between imperfectivity and unmarked verbs was not a strict one (e.g., verbs without prefixes could also have a perfective meaning (Brinton 1988: 200)). Moreover, in many of their 13 In some studies, it is the prepositional construction beon/wesan + in/on/a + V-ing that is identified as the modern progressive s ancestor, yet this prepositional pattern did not occur in progressive contexts before the 14 th century. Therefore, we follow Ziegeler (1999), who suggests that it merged with the independently developing beon/wesan + V-ende construction and, consequently, considerably influenced the functional development of this construction in late Middle and Modern English. 14 In Slavic, perfective aspect is said to signalize the attainment of a final boundary, totality and/or temporal specificity in sequential relations, in contrast with imperfective aspect, which indicates that the situational boundaries, if any, are neglected (Gvozdanović 2012). Therefore, and given the analogy between Old English and (a type of) Slavic aspectual marking, we put (a)telicity and (im)perfectivity on a par for the purposes of our description of aspect in Old English. Yet we do so only in this context, since these aspectual concepts pertain to different levels of analysis and are distinct in (non-slavic) languages in which perfective and imperfective aspect are grammatically expressed, rather than constituting some grammaticalized lexical categories (Dahl 1985: 89), as in Slavic. 17

18 uses the literal (directional or locative) meaning of the prefixes was still transparent, which indicates that these derived verbal markers were not fully grammaticalized as general markers of perfective aspect in the Old English period. Yet, gradually, the literal meanings were bleached, giving way to the expression of aspectual meanings such as perfectivity (Brinton 1988). It is, therefore, safe to say that the Old English aspectual prefix system was relative productive. In Old English, the progressive beon/wesan + V-ende construction already existed, but it appears to have been an unsystematically used stylistic device, rather than a genuine aspect marker. It was used for the expression of imperfectivity and duration (Scheffer 1975; cf. also Bertinetto, Ebert & de Groot 2000 on the durative meaning of the progressive in Old English), to frame another situation, but often also merely to foreground an event in narrative contexts (Núñez-Pertejo 2004: 90-91; Killie 2008; Kranich 2010: 87-88). Various studies (e.g., Scheffer 1975: 162; Núñez-Pertejo 2004: 65-66) indicate that it was most commonly associated with intransitive, durative action verbs and sometimes even with stative predicates, which seems to suggest that it functioned as a concord (or a type-selecting) construction, in the terms of Michaelis (2004; 2011). 15 That is, it typically selected particular verb types with which it was in accordance semantically and simply flagged their duration. In cases where it did occur with non-concordant verbs (non-durative perfective ones), coercion took place and these nondurative verbs were turned into durative ones (i.e., imperfectivized). 16 By the end of the Old English period, the synthetic way of aspect marking by means of prefixes gradually disappeared due to a number of (language-internal and -external) factors (cf. Brinton 1988: 189 for an overview) and, as a consequence, the formal distinction between perfective and imperfective verbs got lost. This evolution appears to have triggered the development of the elaborate system of phrasal verbs still present in contemporary English (cf., e.g., Denison 1985; Brinton 1988). Initially, however, the relevant phrasal particles (up, out etc.) only involved concrete spatial meanings (Brinton 1988: 215), thereby replicating the original function of the prefixes that had (all but) disappeared and not the type of general aspect 15 As Peter Petré (pers. comm.) points out, the stative verbs that allowed a progressive construal in Old English still often involve a component of dynamicity. For example, in The flood flows into the Ocean, an example translated from Old English (Kranich 2010: 86), the overall meaning of the sentence is stative, yet the flowing is in itself dynamic. 16 Michaelis s (2004; 2011) analysis of concord constructions mainly focuses on constructions with a stative meaning that, at the same time, select stative verbs. Hence, we are slightly adapting her analysis by considering atelic duration (rather than pure stativity) as the crucial aspectual parameter that is shared by the construction and the selected verb type. 18

19 marking that these had specialized into at a later stage. 17 McWhorter (2007: 70) further points out that those prefixes that had taken on a more grammatical (perfective) meaning, such as be-, ge- and for-, simply disappeared from the grammar without any substitution (in the form of a particle). In other words, by the 11 th century, the English language was faced with an aspectual vacuum (Strang 1970: ; Núñez-Pertejo 2004: 67), which would constitute an ideal context for the remarkable rise of the progressive in later stages. Data from other Germanic languages provide additional evidence for this association between the loss of aspectual prefixes and the grammaticalization of the progressive. It should be noted that the Proto-Germanic aspectual prefix system was not only adopted into Old English, but also into the older varieties of other West-Germanic languages, such as Dutch and German. In these languages, however, the prefix system has never disappeared and has remained relatively productive (van Kemenade & Los 2003). Ver- (cognate of for-) and be-, for instance, are used to express that the object is fully affected, thus generating a transitive and perfective meaning. Examples of the aspectual use of ver- and be- in present-day Dutch include spelen / verspelen ( play / forfeit ), slapen / (zich) verslapen ( sleep / oversleep ), ademen / beademen ( breath / breath air into ) and zorgen / bezorgen ( take care of / deliver ). Similar oppositions can be attested in German and Old English cf., for instance, German lassen / verlassen and Old English lætan / forlætean as equivalents of Dutch laten / verlaten ( let / abandon ) (van Kemenade & Los 2003: 95). Derivational uses of the perfectivizing prefix gecan also still be attested in contemporary standard Dutch e.g., denken / gedenken ( think / commemorate ) yet it does not seem to be as productive anymore as a perfectivizing device as it used to be in older varieties of Dutch (Van der Horst 2008: ). Most likely this is because its main function nowadays is to mark the past participle, used in, for instance, the perfect construction (which consists of hebben ( have ) or zijn ( be ) followed by a past participle, e.g., Ik heb al ge-slapen ( I have already slept )). Naturally, in this function of ge- its original perfective meaning still shines through and, notably, the formation of the past participle in English does not require prefixation. In sum, present-day Dutch (as well as German) can still, to a certain extent, rely on a system of prefixes to form imperfective/perfective word pairs. Given our hypothesis concerning the link between an aspectual vacuum and the rise of the progressive in English, we would expect that the progressive in Dutch and German is not as 17 Nowadays, in many cases, the literal meaning of verbal particles has faded, making them more suited for the expression of aspectual values. Nevertheless, the system of phrasal verbs can still hardly be claimed to be as grammaticalized as other forms used for the expression of aspect in English, such as the perfect or the progressive constructions. 19

20 grammaticalized as it is in English. This expectation is indeed borne out, as is indicated by, for instance, Ebert (2000). In fact, the only Germanic language in which the progressive seems to have grammaticalized to (about) the same extent as the English progressive is Icelandic (Jóhannsdóttir 2011). Interestingly, and in line with our analysis, Icelandic is the only Germanic language beside English that has fully shed the system of Germanic verbal prefixes (McWhorter 2007: 71). 18 In the Modern period, the progressive which had, during the Middle English period, taken on the be + V-ing form rose remarkably in frequency and underwent a paradigmatic expansion, occurring in contexts that were originally preserved for the simple present (such as the passive). Backgrounded contexts in the past appear to have constituted its main locus of grammaticalization (Petré 2012). The fact that Dutch and German do not seem to have felt the same need to mark past imperfective situations by means of the progressive might not only be due to their relatively productive prefix system, but also to the evolving perfect construction in these languages, which contributed to prevent the rise of an aspectual vacuum by taking up past perfective functions. In Dutch, the perfect had, by the 16 th century, taken up a past perfective meaning, at the expense of the simple past, which only remained unchallenged in past imperfective contexts (Van der Horst 2008: 628). In German, this intrusion of the perfect into the past domain is even more pervasive, as the use of the simple past has become virtually restricted to past imperfective contexts under pressure of the perfect (Boogaart 1999: 156). The English perfect also rose in frequency throughout the ME period up until the 16 th century, yet its rise was then leveled off until the 18 th century, after which it even lost in frequency in American English (Bowie, Wallis & Aarts forthcoming). Importantly, the English present perfect never actually seems to have taken on past perfective meanings. Thus, the English simple past remained aspectually ambiguous throughout, rather than acquiring a (more) imperfective meaning, such that the progressive was free to grammaticalize in past imperfective contexts. Our analysis of course does not a priori exclude that other factors may have influenced the continuing increase in grammaticalization of the English progressive beside the aspectual gap that had arisen in the Middle English period. One such factor is discussed by Petré (2012) and Los & Starren (forthcoming), who point to the general evolution of English from a bounded to 18 In his discussion of the aspectual prefix system in Icelandic, McWhorter (2007: 71) further specifies that the prefixes had actually already been abandoned in Old Norse, which means that they were lacking not only in Icelandic (and Faroese), but also in Mainland Scandinavian languages (Swedish, Norwegian and Danish), yet these languages later borrowed them again from Low German. 20